AMERICAN MUSEUM Novtitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 3270, 22 pp., 9 figures, 5 tables August 10, 1999 Speciation in African Forest Robins (Stiphrornis): Species Limits, Phylogenetic Relationships, and Molecular Biogeography PAMELA BERESFORD! AND JOEL CRACRAFT? ABSTRACT The monotypic genus Stiphrornis (Aves: Turdidae) is revised under a phylogenetic species concept to include four species, one of which, from the southwest Central African Republic, is new. Mitochondrial DNA sequence data are analyzed to explore the phylogenetic relation- ships within Stiphrornis. These data indicate relatively high levels of sequence divergence among the species and corroborate their recognition as diagnosable taxa, a conclusion also supported by morphological evidence. These findings, along with the allopatric distributions of the species, compel attention to their phylogenetic and spatial history, which was not ex- plored when this group was ascribed to a single ‘““biological’’ species. Data reviewed here also suggest that the northwest Congo Basin forest, where the new species was discovered, is more zoogeographically complex than has been previously sus- pected. In addition, application of a phylogenetic species concept emphasizes the narrow en- demism of S. gabonensis and S. sanghensis, along with its implications for conserving their threatened habitats. The findings of this paper also reinforce the notion that patterns of geographic variation in the lowland forests of West and Central Africa are still incompletely understood and that the impact of environmental and geological history on the diversification of the forest avifauna has not yet been fully explored. ' Chapman Graduate Fellow, Department of Ornithology, American Museum of Natural History. ? Curator, Department of Ornithology, American Museum of Natural History. Copyright © American Museum of Natural History 1999 ISSN 0003-0082 / Price $3.80 2 AMERICAN MUSEUM NOVITATES INTRODUCTION During an expedition in 1996 to south- western Central African Republic (CAR) to collect materials for the Hall of Biodiversity in the American Museum of Natural History, specimens of the African Forest Robin (“‘Sti- Phrornis xanthogaster’’) were obtained. Upon studying the material, it became ap- parent that specimens from the Dzanga-San- gha Dense Forest Reserve represented an un- described population. An examination of species limits within the genus was under- taken based on comparisons of museum col- lections in North America and Europe, and taxic interrelationships were investigated by cladistic analysis of mitochondrial DNA (mtDNA) sequence variation. Additional specimens were obtained from the Reserve in June 1998. Biological collections from the center of the Guineo-Congolian forest block (sensu White, 1983)—from southwestern CAR south to northern Congo Republic, and across northwestern Democratic Republic of Congo (DRC; formerly Zaire)—are under- represented in museums, and elements of the forest avifauna have not been subject to in- tensive systematic analysis. Although work- ers have proposed that CAR encompasses a zoogeographic transitional area for other ver- tebrates (Fay, 1988; Joger, 1990), little is known about the patterns of variation in birds within the Congo Basin forest (see also Louette, 1992). For example, there is insuf- ficient data at present to clarify the limits of geographic variation across the region of eastern Cameroon and eastern DRC. Based on patterns of species richness and distribution, Diamond and Hamilton (1980) described areas of endemism for lowland Af- rican passerine birds that were later corrob- orated for both passerine and nonpasserine birds by Crowe and Crowe (1982). These de- scriptions, however, relied on maps in the British Museum atlases of distributions (Hall and Moreau, 1970; Snow, 1978), which re- flect a geographic bias because Hall and Mo- reau (1970) consulted only British and south- ern African museum collections. Thus, spe- cies that Diamond and Hamilton (1980) char- acterized as disjunct across the Congo Basin are now known to occur within it, either be- NO. 3270 cause specimens were represented in other museums (see Louette, 1984) or have been shown to have more extensive ranges due to further collecting and fieldwork (data in Green and Carroll, 1991; Keith et al., 1992; Dowsett and Dowsett-Lemaire, 1997). As a consequence, patterns of species richness and endemism cannot be considered to be well known, especially for taxa in the Congo Ba- sin forest region. At this time, no hypotheses of historical interrelationships among the areas of ende- mism have been proposed based on phylo- genetic analyses; instead, biogeographic in- vestigations have largely focused on inter- montane areas and their relationships or on montane-lowland affinities (e.g., Bruhl, 1997; Fjeldsa and Lovett, 1997; Roy, 1997). While it is to be expected that the expan- sion of a collections-based data set and in- creased fieldwork will clarify distributional information, broadly inclusive “‘biological”’ species concepts have also led to an under- resolution of patterns of geographic varia- tion. This is true for the Stiphrornis group in which some treatments identify patterns of variation in terms of a single taxonomic en- tity (e.g., Map 145 in Hall and Moreau, 1970), thus effectively obscuring the evolu- tionary diversity within the genus, and ob- viating the need to search for patterns of en- demism and clarify historical area-relation- ships when they do, in fact, exist. ACKNOWLEDGMENTS We are grateful to R.W. Dickerman (Uni- versity of New Mexico; AMNH) who col- lected specimens and offered his advice and insight during the study. We would like to thank the following individuals for access to the collections at their museums or for loans of material: C. Erard and E. Pasquet, Musé- um National d’Histoire Naturelle de Paris; M. Louette, Musée Royale de Il’ Afrique Cen- trale; R. Prys-Jones, Natural History Muse- um (London), Department of Ornithology (Tring); J. Bates and D. Willard, Field Mu- seum of Natural History (Chicago); David Allan, Durban Natural Science Museum; and K. Garrett, Natural History Museum of Los Angeles County. R. Dekker, Nationaal Na- tuurhistorisch Museum (Leiden) provided in- 1999 formation about Stiphrornis in that museum’s collection. M. LeCroy (AMNH), R. Prys- Jones, and E. Warr (Natural History Muse- um) provided assistance with literature. In the CAR and in New York, the AMNH Ex- hibition Department provided support. The Ministére des Eaux, Foréts, Chasses, Péches et du Tourisme gave us permission to collect in the CAR. Richard Carroll, Allard Blom, Tony Mokombo, and Bernard Difara of the World Wildlife Fund U.S. and Urbain Nga- toua, National Director of the Dzanga-San- gha Conservation Project, greatly facilitated our work in the CAR as did many local res- idents of Bayanga. D. Lunde, P. Sweet, and A. Porzecanski (AMNH) provided invalu- able assistance in the field. J. Feinstein and J. Groth (AMNH) provided primers and ad- vice on the molecular analysis. We would like to thank Dennis Finnin for photography. The manuscript benefitted from discussions with G. Barrowclough, S. Keith, and the comments of two anonymous reviewers. The 1996 expedition to the CAR was financed by the AMNH Biodiversity Hall project and by a grant from the National Science Founda- tion (ESI-9453023). The June 1998 expedi- tion was sponsored by a grant from the Cen- ter for Biodiversity and Conservation at the AMNH, and we are grateful to Dr. Francesca Grifo, Director of the Center, for her advice and support. Molecular work was funded in part by grants from the City College (City University of New York) Dissertation Grant Program, and the Frank M. Chapman Me- morial Fund (to PB). This work was also supported in part by the Leonard J. Sanford Fund in the Department of Ornithology. The molecular research reported in this paper is a contribution from the Lewis B. and Doro- thy Cullman Research Facility at the Amer- ican Museum of Natural History and has re- ceived generous support from the Lewis B. and Dorothy Cullman Program for Molecular Systematics Studies, a joint initiative of the New York Botanical Garden and the Amer- ican Museum of Natural History. MATERIALS AND METHODS Museum study skins were examined for evidence of variation and diagnosability in features of external morphology. Mitochon- BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 3 drial DNA (mtDNA) sequence data for the entire cytochrome-b gene (1143 base pairs) were obtained for 13 Stiphrornis individuals plus two individuals from an outgroup taxon, Sheppardia cyornithopsis. Vocalizations were analyzed with Canary 1.2 (Charif et al., 1995). MATERIALS EXAMINED We examined 290 specimens from collec- tions at the AMNH, the Field Museum of Natural History (FMNH), the Natural His- tory Museum of Los Angeles County (LACM), the Natural History Museum (BMNH), the Muséum National d’ Histoire Naturelle de Paris (MNHN), the Musée Roy- ale de l’Afrique Centrale (MRAC), and the Durban Museum. Specimens examined from each of the four taxa included (6 = male, ? = female, ?=sex undetermined): S. erythro- thorax (AMNH 1264, 82; MRAC 84, 28, 2?; MNHN 12; BMNH 3¢ 38 17), S. ga- bonensis (AMNH 166, 102; DM 3<, 22; MRAC 16, 12, 2?; MNHN 4¢é, 52, 37; BMNH 864, 62, 6?), S. xanthogaster (AMNH 3764, 212, 4?; DM 56, 22; LACM 83,42; FMNH 106,42; MRAC 384, 128, 5?; MNHN 34, 22, 1?; BMNH 192, 8B), and S. sanghensis (AMNH 5¢é, 102, 8?). These specimens represented 99 localities across the range of Stiphrornis in the forests of Africa, including Liberia, Sierra Leone, Ghana, Nigeria, Cameroon, Gabon, Equato- rial Guinea, the Republic of Congo, the Con- go Democratic Republic, Sudan, Uganda, and Kenya; localities with unambiguous co- ordinates are presented in the appendix. MOLECULAR METHODS Collecting locales and other information pertaining to the 15 individuals sampled for the molecular analysis are presented in table 1. Genomic DNA was extracted from small pieces of tissue by boiling in 5% (w/v) Chel- ex (Bio-Rad, Hercules, CA) solution. Target regions (see table 2 for primers) of the cy- tochrome b gene were first obtained with 10 wwL PCR reactions (1 wL DNA, | wL each of 10 pM L/H primers, | wL of 2mM dDNTPs, 0.15 pL Tag polymerase [Promega, Madi- son, WI] and 2 wL buffer), placed in an Ida- ho Technologies air thermocycler at the fol- Stiphrornis erythrothorax 1 . erythrothorax 2 NHNHRHRHRHHRRHARARAARN Name gabonensis 1 . gabonensis 2 . gabonensis 3 sanghensis | sanghensis 2 . sanghensis 3 . sanghensis 4 . sanghensis 5 . xanthogaster 1 . xanthogaster 2 . xanthogaster 3 AMERICAN MUSEUM NOVITATES Museum AMNH 827588 AMNH 827589 MNHN 1998-786 MNHN EP1-50 MNHN EP1-64 AMNH 831845 AMNH 831847 AMNH 10836 AMNH 24731 AMNH 831846 FMNH 357243 FMNH 357245 FMNH 357244 TABLE 1 Information on Specimens Used in Molecular Analysis AMNH 827632 AMNH 827650 Sheppardia cyornithopsis 1 S. cyornithopsis 2 lowing reaction conditions: denaturation at 94°C for 5 min, annealing at 50°C for 2 min, and extension at 71°C for 20 sec, all for 40 cycles. 5 wL of these PCR products were run out in 2% low-melting-point agarose gels, vi- sualized with ethidium bromide and ultravi- olet light, and the relevant sized bands ex- cised, diluted in 190uL H,0 and melted at 72°C for 25 min. 40 pL PCR reactions were prepared with 1.5 wL of the gel-purified products, 2 wL each of the 10 wM L/H prim- ers, 4 pL 2mM dDNTPs, 8 pL buffer and 0.2 wL Tag polymerase and cycled in the air thermocycler under the following conditions: denaturation at 94°C for 8 min, annealing at 55°C for 8 min, and extension at 70°C for 22 sec. These products were purified with the NO. 3270 Genbank accession Locality no. near Ziggida, Lofa County, Liberia AF 136724 near Ziggida, Lofa County, Liberia AF136725 Nditam, Cameroon (5°21'N, 11°14’E) AF136726 Nditam, Cameroon AF136727 Nditam, Cameroon AF136728 Dzanga-Sangha Reserve, CAR AF136729 Dzanga-Sangha Reserve, CAR AF136730 Dzanga-Sangha Reserve, CAR AF136731 Dzanga-Sangha Reserve, CAR AF136732 Dzanga-Sangha Reserve, CAR AF136733 Ituri Forest, DRC AF136734 Ituri Forest, DRC AF136735 Ituri Forest, DRC AF136736 near Ziggida, Lofa County, Liberia AF136722 near Ziggida, Lofa County, Liberia AF136723 Gene Clean II system (BIO 101, Inc., San Diego, CA) resulting in a final suspension of DNA in 18 pL HO. 2.5 pL of this purified DNA was prepared for asymmetric cycle se- quencing with 1.5 wL 10mM primer and 3 wL of dRhod Terminator RR Mix (Perkin EI- mer) and cycled in a Perkin Elmer 9600 at the following reaction conditions: initial de- naturation at 96°C for | min, followed by 32 cycles of denaturation at 96°C for 10 sec, an- nealing at 50°C for 5 sec, and extension at 60°C for 3 min. These products were filtered through sephadex columns (Princeton Sepa- rations), dried for 30 min in a Speed-Vac (Savant, Hicksville, NY), and suspended in a Blue Dextran-150 mM EDTA:formamide loading buffer. Samples were run out on 5% TABLE 2 Primer Information Primer numbers follow those of the Gallus sequence (Desjardins and Morais, 1990). Primers are arranged according to their paired use in initial amplifications; L and H refer to light and heavy strands, respectively. L14578 (NDS) 5-ctaggaatcatcctagccctaga-3 H15104 5-tgtgtcagecatattggacgtctcggc-3 L15068 5-actagcaatacactacacagcaga-3 H15505 5-tgcatgaattcctattgggttgtttgatcc-3 L15236 5-ttcctatacaaagaaacctgaaa-3 H15710 5-atagcgtaggcgaataggaagtatc-3 L15656 5-ccagacctcctaggagacccaga-3 H16065 5-aacgcagtcatctccggtttacaagac-3 1999 Long Ranger (FMC, Philadelphia, PA) gels in TBE buffer in an Applied Biosystems, Inc. (Foster City, CA) 377 automated sequencer. Sequences have been deposited in GEN- BANK with accession numbers AF136722- AF 136736. SYSTEMATIC METHODS Sequences were assembled and aligned with Sequencher 3.0 (Gene Codes Corpora- tion, Ann Arbor, MI) and verified by eye. Because the sequences are from coding genes, alignment was straight-forward. For phylogenetic analysis, most-parsimonious trees were obtained through a branch-and- bound search using PAUP* (Swofford, 1998). Sequences from two individuals of Sheppardia cyornithopsis were used to root trees. Sequence divergence measures were estimated for all pairwise comparisons both as uncorrected (p) distances and under the HKY85 model as implemented in PAUP*, with base frequencies and transition/trans- version substitution rates based on observed values. SYSTEMATICS OF STIPHRORNIS Although there are no phylogenetic hy- potheses at present about the relationships of Stiphrornis to other Turdidae, taxonomists have broadly concurred in placing Stiphror- nis in a group of “African Robins”? (White, 1962) or “African forest-dwelling robins”’ (Irwin and Clancey, 1974) along with Shep- pardia, Pogonocichla, Swynnertonia, Cos- sypha, and Alethe; the exact membership of this assemblage varies, and some authors as- sign certain genera to the chats within the Muscicapidae (e.g., Jensen, 1989; Sibley and Monroe, 1990). Sharpe (1903) recognized Stiphrornis as a polytypic taxon, but within a few decades W. L. Sclater’s Systema Avium Aethiopicarum (1930) ranked all Stiphrornis species as sub- species within a single species. Although the compilers of ‘“‘Peters’ Checklist’? (Mayr and Paynter, 1964) placed Stiphrornis within Er- ithacus, most ornithologists continued to maintain Stiphrornis (White, 1962; Hall and Moreau, 1970), and Irwin and Clancey (1974: 7) explicitly objected to the expansion eos of Erithacus to include African forms ‘“‘in BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 5 view of the manifest differences between the African forest-dwelling robins [and] the type of the genus Erithacus.” White (1962) recognized three subspecies: Stiphrornis erythrothorax erythrothorax, S.e. gabonensis, and S.e. xanthogaster. Whereas geographic variation in Stiphrornis has his- torically been recognized at some taxonomic level, Hall and Moreau (1970) mapped Szti- Phrornis as a monotypic taxon in their dis- tributional atlas. In the following revision we recognize four phylogenetic species within Stiphrornis. Genus Stiphrornis Hartlaub 1855 TYPE SPECIES: Stiphrornis erythrothorax Dabocrom, Ghana (Nationaal Natuurhisto- risch Museum [Leiden], not accessioned). DIAGNOsIS: The genus is characterized by a narrow, lateromedially compressed _ bill, short legs and tail, and a white loral spot. Compared to other African forest turdines, Stiphrornis differs in the color pattern of the chin, throat, and breast, which is not distrib- uted further down onto the sides of the belly and flanks as in other taxa, but is instead markedly demarcated between the breast and belly plumage. Within the genus, plumage and color patterns provide diagnostic char- acters for each species (Table 3). INCLUDED SPECIES: Stiphrornis erythro- thorax, S. gabonensis, S. xanthogaster, S. sanghensis n. sp. DISCUSSION: In the original description of the genus, Hartlaub’s (1855) discussion of the type species S. erythrothorax was fol- lowed by one for “S. superciliaris,”’ a form that had been named Sylvia prasina and has been recognized as Hylia prasina (Sylviidae) since 1859 (Sharpe, 1883). In 1874, Rei- chenow described ‘“‘Stiphrornis albotermi- nata,’’ a taxon already described as a nectar- iniid, Nectarinia gabonica, and since 1930 has been named Anthreptes gabonicus (Chapin, 1954). Stiphrornis erythrothorax Hartlaub, 1855 Figure | HOLOTYPE: Male from Dabocrom, Ghana; type in Nationaal Natuurhistorisch Museum (Leiden), not accessioned. DIAGNOsIs: The only species of the genus AMERICAN MUSEUM NOVITATES NO. 3270 ponen- Teese aA Fer 1999 with an olive-green wash on all upper sur- faces. S. erythrothorax shares with S. gabo- nensis a russet throat and chin, and white feathers on the belly. DISCUSSION: Stiphrornis erythrothorax ranges from Sierra Leone to the Niger Delta (e.g., Degema, Nigeria: AMNH 599790; see fig. 4), primarily in lowland forest but also in riverine and savanna habitats within the Upper Guinea region (Keith et al., 1992). Stiphrornis gabonensis Sharpe, 1883 Figure | HoLotTyPe: An adult (indeterminate sex) collected in ‘‘Gaboon,’”’ BMNH 1876.5.23. 206. DIAGNOsIs: Chin and throat russet and bel- ly white, like S. erythrothorax, but differs in having dorsal plumage dark or slaty gray with a faint olive wash, not as green as in S. erythrothorax. DISCUSSION: S. gabonensis is limited to ev- ergreen coastal forest, ranging from just east of the Niger River delta (records from Kum- ba and Mamfe, Nigeria) south to Gabon (and possibly limited by the Ogooué River; e.g., Kango, Gabon: AMNH 345033) east to Mal- en, Cameroon (e.g., AMNH 800617). This species is also found on the island of Bioko, Equatorial Guinea. Sharpe (1883) named this species upon comparing specimens of Stiphrornis while compiling volume 7 of the British Museum Catalogue. The new taxon was described as “very similar’ to S. erythrothorax but dis- tinguished by the dorsal plumage being ‘dark slaty grey with a faint olive tinge.” < Fig. 1. BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS vi Stiphrornis xanthogaster Sharpe, 1903 Figure | Ho.LortyPe: Adult Gndeterminate sex) from the River “Ja” (Dja), Cameroon, BMNH 1903.7.16.100. An immature syntype (inde- terminate sex) is also in the collection. D1AGNosIs: Differs from other Stiphrornis in having chin and throat tawny and abdom- inal feathers pale cream. The dorsal plumage is gray-brown lightly washed olive, not as dark gray as in gabonensis and not green as in erythrothorax. DISCUSSION: S$. xanthogaster is the most widespread member of the genus, ranging from the River Dja in Cameroon (e.g., AMNH 599798) south through northeastern Gabon (e.g., MNHN 517, Bélinga) to Lu- kolela, DRC (e.g., AMNH 2696983) east through northeastern CAR (e.g., LACM 84941), Sudan, the Ituri forest, and Uganda to western Kenya (e.g., Kipkabus, AMNH 788638) (fig. 4). Although mainly restricted to lowland primary forest, in eastern DRC S. xanthogaster occurs in transitional forest up to 1400 m (Keith et al., 1992). S. mabirae was described by Jackson (1910) based on specimens collected in the Mabira Forest, Uganda, and diagnosed by the chin being as dark as the rest of the throat region and by the upperparts being “‘more olive.”’ Chapin (1953) did not find it “easy to distinguish specimens of xanthogaster ... from those of mabirae’’ and White (1962) subsumed mabirae into xanthogaster. The distribution of the relative intensity of chin and throat pigmentation and the saturation of the dorsal olive wash varies among individ- uals, not geographically, thus precluding the use of these characters to delineate mabirae Ventral and dorsal views of Stiphrornis taxa. Left to right, top: S. erythrothorax, AMNH 825486 (Liberia) and AMNH 827586 (Liberia); S. gabonensis, AMNH 297701 (Bioko, Equatorial Guin- ea) and AMNH 800617 (Cameroon); bottom: S. xanthogaster, AMNH 161024 (DRC) and AMNH 164023 (DRC); S. sanghensis, AMNH 832117 (CAR) and AMNH 832121 (CAR). Photograph by D. Finnin. Fig. 2. Adult male Stiphrornis sanghensis, photographed in in Dzanga-Sangha Dense Forest Re- serve, CAR, June 1998. Photograph by J. Cracraft. Fig. 3. Juvenile male Stiphrornis sanghensis, photographed in in Dzanga-Sangha Dense Forest Re- serve, CAR, June 1998. Photograph by J. Cracraft. 8 AMERICAN MUSEUM NOVITATES as a distinct taxon. Specimens of S. xantho- gaster from the Kivu region in DRC at Mu- sée Royale de l’Afrique Centrale and the Durban Museum have a more pale yellow wash to the belly feathers than do specimens from other parts of the range. Stiphrornis sanghensis, new species Figures 1-3 HoLotyPe: Adult male, AMNH 832121, collected in the Dzanga-Sangha Dense Forest Reserve (2°55’N, 16°15’E, ca. 1 km north of Bayanga, Sangha-Mbaéré Prefecture), Cen- tral African Republic, on 13 June 1998, by P. Beresford. PARATYPES: AMNH 832123, adult female, 6 June 1998; AMNH 832126, adult female, 13 June 1998; AMNH 832120, adult male, 14 June 1998; AMNH 832124, adult female, 17 June 1998; AMNH 832128, juvenile (in- determinate sex), 19 June 1998; AMNH 832127, immature female, 2 July 1998. The following were collected on the west bank of the Sangha River, across from the previous locality: AMNH 832125, adult female, 24 June 1998; AMNH 832116, adult female, 24 June 1998; AMNH 832117, adult (indeter- minate sex), 24 June 1998; AMNH 832122, adult female, 24 June 1998; AMNH 832118, adult male, 24 June 1998; and AMNH 832119, adult female, 25 June 1998 (pre- pared as flat skin and partial skeleton). The following were collected at the confluence of the Babongo and Sangha Rivers (2°59'N, 16°14’E, ca. 8 km north of Bayanga): AMNH 831845, adult (indeterminate sex), 16 November 1996; AMNH 831846, sub- adult (with ossified skull, indeterminate sex), 17 November 1996; AMNH 831847, adult male, 24 November 1996; AMNH 831848, adult (andeterminate sex), 3 December 1996. Skeletons: AMNH 24732, sex undetermined; AMNH 24731, male; AMNH 24871, female; AMNH 24869, male; AMNH 24870, adult female (prepared as flat skin and partial skel- eton). Fluid preserved: AMNH_ 10836, AMNH 10863. DIAGNOSIS: Distinguished from its conge- ners by a deep orange-yellow chin, throat, and upper breast, and a yellow wash to the belly feathers. DESCRIPTION OF HOLOTYPE: Crown and NO. 3270 forehead dark slate, lightly tinged olive; nape and upper back basally gray, feathers tipped dark olive green; lower back and upper tail coverts basally gray, tipped with lighter olive green; white loral spot, black malar feathers, auriculars dark blue-gray; chin, throat, and breast bright orange-yellow, appearing iri- descent at certain angles; feathers at sides of breast edged dark gray; upper and lower bel- ly feathers basally dark gray becoming cream and tipped yellow; feathers of flanks and sides more gray; remiges brown with leading edge washed olive; rump, upper tail coverts, and rectrices gray-brown with yellow-green wash on dorsal surface. See Table 4 for mea- surements of the type series. DESCRIPTION OF PARATYPES: Two immature individuals were collected. On the younger bird (AMNH 832128), the chin, throat, upper breast, crown, and dorsal feathers all bear subterminal dull orange spots, with the belly feathers basally gray and tipped white, pro- ducing a mottled effect. The primaries are gray-brown edged olive. The immature tail feathers are pure russet, and a set of emerg- ing rectrices are gray-brown. The other im- mature bird, a male (AMNH 832127, fig. 3), bears a plumage intermediate between the younger and adult plumages: only the upper breast and a few dorsal feathers bear the dull orange subterminal spot, while both the chin and most of the belly (except for a thin, cen- tral line of mottled gray and white feathers) are lemon yellow. The primaries are gray- brown edged olive, and the rectrices are gray brown with an olive wash on the dorsal sur- faces. A subadult paratype (AMNH 831846), which is molting to adult plumage, shows more extensive basal gray on the (paler) bel- ly and flank feathers and bears rufous-edged primaries. An adult male not in breeding condition (AMNH 832120) has the crown and nape not as dark, with feathers tipped paler olive green. EtyMo_Locy: The name refers to the type locality, the Dzanga-Sangha Dense Forest Reserve in the Sangha-Mbaéré Prefecture of the Central African Republic. The reserve lies at the northern edge of the Guineo-Con- golian forest block of White (1983) and is dominated by mixed semi-deciduous ever- green rainforest (see Green and Carroll, 1991). We propose the English name ‘“‘San- 1999 i pe) re S. erythrothorax Z NO Sra ‘A S. gabonensis BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 9 _ S. xanthogaste fe & C? Fig. 4. Distributions of species of Stiphrornis (after Keith et al. (1992) and material examined in this study). Range limits of taxa in the western Congo Basin are poorly known at present, therefore it is uncertain whether parapatry exists between S. gabonensis and S. erythrothorax. gha Forest Robin’ in reference to the type locality. DISCUSSION: To date, S. sanghensis is known only from the Dzanga-Sangha Dense Forest Reserve, where it was commonly ob- served and captured in the lower strata of primary forest, old second-growth forest, and moderately inundated forest along both sides of the Sangha River. Future collecting of Sti- Phrornis from eastern Cameroon, northern Congo Republic, and northwestern DRC should clarify the range limits of S. sangh- ensis, S. xanthogaster, and S. gabonensis. Birds collected in June 1998 were in breeding condition. Two females had yolking eggs low in their oviducts, another an en- larged oviduct. Three males had large testes, measuring between 7 X 5 mm and 10 X 5 mm, compared to testes sizes of 1.5 X 2 mm in the nonbreeding male collected in Novem- ber 1996. Vocalizations were heard throughout the day during June and July 1998. Two types of calls were heard (fig. 5) and verified by play- back; both calls were recorded in the field in 1998 on three days from at least two indi- viduals. ““Type A’ begins with a few high chirps and continues with a series of modu- lated notes; several of these phrases may be given consecutively for several minutes in an unbroken stream. The Type A call was also heard as single or paired phrases, especially when answered by another Type A call from a different individual. The ““Type B”’ call is 10 AMERICAN MUSEUM NOVITATES Type A kHz pr nee ee gs neal hee ee ny 1.0 2.0 3.0 4.0 5.0 6.0 kHz | J | | hf : 1.5 Hs, © 5 ae castbonn CoP, id AS So RECLA. ven a. cael | ee 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Seconds Fig. 5. Forest Reserve, CAR, June 1998. NO. 3270 Spectrogram of two types of Stiphrornis sanghensis calls, recorded in Dzanga-Sangha Dense 1999 BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 11 6. " A S. sanghensis 4.5 4.0 3.5 3.0 a “| a 0.0 —S ee Z| | Mik ih nere ak ohh Aste "teen 6! Sem aka | eg Seconds 6.0 | BS. xanthogaster 55 5.0 ; | 45 ) " fi ae My 40 | | a »/! nT S| , 3 i Prtyid | roe F wilh I a Wi re i ih i i | 2.5 : 2.0 4 154 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Seconds Fig. 6. Spectrograms of two Type A Stiphrornis calls. (A) S. sanghensis, recorded in Dzanga-Sangha Dense Forest Reserve, CAR, June 1998; (B) S. xanthogaster, recorded near Mt. Hoyo, DRC, Cornell Library of Natural Sounds Catalogue Number LNSO1456. 12 AMERICAN MUSEUM NOVITATES best described as a rolling trill, with no clear demarcation of units or phrases, and was also given for continuous bouts of up to several minutes’ duration. The duetting behavior de- scribed by Brosset and Erard (1986) was not heard in CAR. One Type A call from S. xanthogaster (Mt. Hoyo, 1°20’'N, 29°46’E; Cornell Library of Natural Sounds catalogue number LNSO1456) was compared to that of S. sanghensis (fig. 6). The three opening notes are structurally different, and the remaining parts of the phrases are modulated differently between the two individuals. Three Type B calls are compared in figure 7, although the homology of these calls has not been estab- lished. The spectrogram of the call from the individual from eastern DRC (Lolwa, 1°23'N, 29°30’E; Cornell Library of Natural Sounds catalog number LNSO1091) repre- sents a phrase that is repeated continuously; by contrast, the call of S. sanghensis seems to have no distinct phrase markings. Similar- ly, the call from the individual from Uganda (Keith and Gunn, 1971) has distinct phrases, of which two are shown in the spectrogram. The calls of the two S. xanthogaster individ- uals are as different from each other as each of them are from S. sanghensis. Until more well-documented vocalizations are studied, it is impossible to determine which of these differences reflect variation among individ- uals, populations, or species. At present, it appears that more field recordings, made at various times of the year and in various be- havioral contexts, will be required in order to describe the vocal repertoire of Stiphrornis species. Although vocalizations of S. ery- throthorax and S. gabonensis were not ana- lyzed in this study, they are onomatopoei- cally described as different in Keith et al. (1992), and C. Chappuis (personal commun.) finds them to differ markedly from each oth- er and from those of S. xanthogaster. As noted by Keith et al. (1992), adult fe- males of S. erythrothorax have duller throats than do adult males, and adult females of S. gabonensis have a slight olive wash on the dorsal feathers. Males of S. sanghensis show more intense pigmentation on the throat and upper breast, and the crown and face are of a deeper black than in females. No sexual dimorphism was reported for S. xanthogaster NO. 3270 by Keith et al. (1992). The dichromatism in S. gabonensis may explain the “greener back”’ (Louette, 1981: 128) of unsexed S. ga- bonensis specimens from Ngoumé, Came- roon (5°30'N, 11°26’E) as well as the appar- ent “‘intergradation’’ in western Cameroon mentioned by Hall and Moreau (1970: 123). Since immature specimens of S. sanghensis have paler chins and may thus resemble S. xanthogaster specimens, comparative analys- es must account for differences among indi- viduals of different age and breeding condi- tion. Similarly, the yellow wash on the belly feathers that is one of the diagnostic features for mature S. sanghensis is more sulphurous than the cream to pale yellow tint on the bel- lies of some specimens of S. xanthogaster, a condition also seen on the belly of a subadult S. sanghensis (AMNH 831846). RELATIONSHIPS AND BIOGEOGRAPHIC PATTERNS OF STIPHRORNIS CLADISTIC ANALYSIS Sequence analysis and comparison of complete cytochrome b sequences for indi- viduals of 13 Stiphrornis and 2 Sheppardia demonstrated that each represented a sepa- rate mitochondrial haplotype: thus they were all included in a cladistic analysis. Five equally parsimonious trees of 243 steps were found using 183 parsimony-informative characters, and the strict consensus of these trees is shown in figure 8. The individual haplotypes assigned to each species are seen to cluster together, each cluster having a bootstrap value of 100%. The new species, S. sanghensis, is postulated to be the sister species of S. xanthogaster and the two form one lineage within Stiphrornis. The other species, S. erythrothorax and S. gabonensis, are themselves united and form the sister group of S. sanghensis and S. xanthogaster. Within Stiphrornis, 125 sites were variable (10.9%); changes at third positions com- prised most of this variation (104 sites), fol- lowed by changes at first positions (18) and then 3 changes at second positions. The tran- sition: transversion ratio was 7:1. A phylogram in figure 9, representing one of the five equally parsimonious trees, is used to illustrate the relative branch lengths 1999 fe a i = Figy “2. BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 13 604 A §. sanghensis 5.5 SS ae a ge Fe RS Ss ge gh UR gS 0.0 0.2 04 0.6 0.3 6.0 , ' BS. xanthogaster 5.5 5.0 4.8 ee ee a a once 3.5 3.0 Zs Ven eches Seee bea ia 2.0 + 1.5 2 ene a og 14 1.6 1.8 20 .| i | | e | RY | | | | B | i 0.8 1.0 Seconds 604 CS. xanthogaster a 5.5 = 5.0 - SS ET 0.0 0.2 0.4 0.6 0.8 10 12 14 1.6 1.8 2.0 Seconds Spectrograms of three Type B Stiphrornis call phrases. (A) S. sanghensis, recorded in Dzan- ga-Sangha Dense Forest Reserve, CAR, June 1998; (B) S. xanthogaster, recorded near Lolwa, DRC, Cornell Library of Natural Sounds Catalogue Number LNSO1091; (C) S. xanthogaster, Bwamba Forest, Uganda (Keith and Gunn (1971)). 14 AMERICAN MUSEUM NOVITATES NO. 3270 100 S. erythrothoraxl Liberia S. erythrothorax2 69 yz S. gabonensis! Nditam, 100 S. gabonensis2 | Cameroon S. gabonensis3 S. sanghensis2 Dzanga-Sangha 100 F af Reserve, CAR S. sanghensis3 S. sanghensis4 100 100 ) S. sanghensis1 S. sanghensis5 S. xanthogasterl 100 | Ituri Forest, S. xanthogaster2 | DRC S. xanthogaster3 Sheppardia cyornithopsis | Sheppardia cyornithopsis2 Fig. 8. Strict consensus of five most-parsimonious trees (243 steps, CI=0.868) derived from a branch-and-bound search on cytochrome-b sequence data for 13 Stiphrornis haplotypes and two out- group haplotypes. Bootstrap values based on 500 replicates. => Fig. 9. One of five most-parsimonious trees obtained from the analysis of figure 8 (branch-and- bound search (DELTRAN optimization), derived from parsimony analysis of cytochrome-b sequence data for 13 Stiphrornis haplotypes and two outgroup haplotypes, chosen to illustrate relative branch lengths. 1999 BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS Za. 8. erythrothorax1 30 1S. erythrothorax2 2 S. gabonensis\ S. gabonensis2 25 ue S. gabonensis3 1! §. sanghensis1 98 1 oes: sanghensis5 1S. sanghensis2 Je S. sanghensis4 3S. sanghensis3 18 S. xanthogaster! 1 es S. xanthogaster2 S. xanthogaster3 Sheppardia cyornithopsis1 Sheppardia cyornithopsis2 16 Plumage/body part Forehead, forecrown, and crown Nape, mantle, and back Chin, throat and breast Upper and lower belly Upper wing coverts Rump, upper tail co- verts and dorsal sur- face of rectrices Lesser underwing co- verts Flank and tibiotarsus AMERICAN MUSEUM NOVITATES TABLE 3 Summary of Plumage Variation in Stiphrornis sanghensis Gray to dark gray with olive wash Gray with olive wash Bright yellow-or- ange Yellow Dark brown Gray washed yel- low-green Gray tipped pale yellow Light gray tipped yellow xanthogaster Gray to dark gray with olive wash Gray with olive wash Tawny; varies to pale beige at chin and throat Cream Dark brown Gray with olive wash generally brighter than on dorsals Gray tipped cream Light gray tipped cream gabonensis Darker gray with faint olive wash Slaty gray, only faintly tinged ol- ive Russet White Dark brown to slaty gray Gray with olive wash generally brighter than on dorsals Gray tipped white Light gray tipped white NO. 3270 erythrothorax Gray with green wash Gray with green wash Russet White Dark brown edged olive Gray with green wash, not distinct from dorsals Gray tipped white Light gray tipped white TABLE 4 Measurements of Stiphrornis sanghensis Type Series Weight is given in grams. Measurements (in millimeters) were taken as follows: wing from carpal to tip of longest primary (chord); tail from pygostyle to tip of longest rectrix; and culmen from tip to base of skull. AMNH Weight Wing Tail Culmen_ Tarsus 831845 17.1 64 37 14.4 D2, 831846 18.1 56 32 1229 21.6 831847 18.0 62 35 14.3 23.2. 831848 18.0 67 39 1553 22-3 832116 15.5 60 31 14.8 20.2 832117 17.0 62 36 12.6 22.6 832118 16.5 66 37 14.6 22.6 832120 175 65 37 13.9 19.1 832121 £75 64 38 15.2 21.4 832122 15-0 59 32 12.9 21.4 832123 22.5 61 34 14.4 19.8 832124 16.0 59 30 13.9 21.9 832125 155 5A, 33 14.5 21.4 832126 17.0 60 36 14.1 22.9 832127 16.0 62 31 11.6 20.6 832128 15:5 53 16 14.0 20.8 among the haplotypes and the five taxa. The mean uncorrected pairwise divergence values (table 5) between erythrothorax and either xanthogaster or sanghensis is 5.8%, whereas gabonensis differs from the other two by 6.0% sequence divergence. Although rate homogeneity within cytochrome Db has not been demonstrated for passerine birds, these values are comparable to distances seen among species in other groups (e.g., Hackett, 1996; Klicka and Zink, 1997; Avise and Walker, 1998). These data are also consistent with the hypothesis that S. xanthogaster and S. sanghensis have diverged relatively more recently from each other, with a mean pair- wise divergence value of 2.8% between them, than have S. gabonensis and S. ery- throthorax which exhibit 5.4% divergence between them. Based on the distribution of plumage fea- tures among African turdines, the russet chin, throat, and breast of S. erythrothorax and S. gabonensis appears to be the primitive con- dition, whereas the lighter condition shown 1999 by S. xanthogaster and S. sanghensis is prob- ably a shared derived character. This inter- pretation is congruent with the phylogenetic hypothesis implied by the molecular data. MOLECULAR BIOGEOGRAPHY The distributional information, summa- rized under each species description above, delineates specific ranges for S. erythrothor- ax (referred to here as the Upper Guinea low- land forest region, limited in the east by the Cameroon highlands) and S. gabonensis (coastal evergreen forests of the Cameroon- Gabon region); the ranges of S. sanghensis and S. xanthogaster are less clear. Both spe- cies occur in the Congo Basin, but further research is required to describe their distri- butions with more precision. Current distri- butional information for S. gabonensis shows it to be more habitat-specific—trestricted to the coastal evergreen forests of the Came- roon-Gabon region by drier habitats—than are the other Stiphrornis taxa that appear to frequent several habitat types. S. erythro- thorax appears to tolerate a wide range of forest types in the lowlands of the Upper Guinea region and ranges east to Mt. Cam- eroon and the highlands of the Adamawa Plateau. The area relationships revealed by the phylogenetic results indicate that the Upper Guinea and Cameroon-Gabon regions are more closely related to each other than either is to areas in the Congo Basin. Assuming the history of Stiphrornis mirrors that of the ar- eas, the relative branch lengths on the phy- logram (fig. 9) suggest that these areas have been isolated from each other longer than have the areas within the Congo Basin forest. DISCUSSION Phylogenetic relationships based on nucle- otide characters among the Stiphrornis hap- lotypes corroborate the recognition of four distinct phylogenetic species. The branch lengths (fig. 9) suggest that each clade has been isolated for a relatively long time (see below). Thus, the historical independence of these taxa indicated by this analysis com- bined with their apparently allopatric distri- butions may also imply some ecological dis- tinctness as well, differences that might not TABLE 5 Pairwise Distances Above diagonal, uncorrected pairwise distances; below diagonal, number of bases differing in pairwise comparisons. 15 14 13 #2 11 10 0.003 0.108 0.105 0.113 O.412 O.112 O.114 O.1115 0.114 O.114 O15 O.111 O11 0.109 0.104 0.107 0.06 0.06 1 Sheppardia cyornithopsis 1 2 S. cyornithopsis 2 0.108 0.061 0.109 0.109 O.113 0.114 O.113 0.113 O.1115 0.108 0.059 0.060 0.052 0.110 0.060 0.059 0.059 0.061 0.052 0.059 122 3 Stiphrornis erythrothorax 1 123 0.059 0.059 0.060 0.059 0.059 0.061 0.062 0.061 0.052 120 4 S. erythrothorax 2 5 S. gabonensis 1 0.061 0.060 0.059 0.059 0.061 0.057 0.056 0.055 0.014 60 59 67 60 59 68 126 125 125 129 128 128 0.058 0.058 0.059 0.057 0.056 0.055 0.06 0.059 6 S. gabonensis 2 0.059 0.064 0.066 0.061 0.064 14 68 16 70 69 7 S. gabnonensis 3 8 S. sanghensis 1 9 S. sanghensis 2 10 S. sanghensis 3 11 S. sanghensis 4 12 S. sanghensis 5 BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 0.002 0.002 0.030 0.029 0.028 0.003 75 129 68 68 130 131 0.029 0.028 0.027 0.003 74 a 67 69 69 130 0.030 0.029 0.028 0.003 68 68 68 66 66 68 129 130 0.026 0.027 73 129 68 68 68 | 130 AS 70 65 70 cA 70 69 132 34 33 32 oo oi 34 33 32 33 a2 31 34 33 32 70 69 65 124 123 bay 13 S. xanthogaster | 14 S. xanthogaster 2 15 S. xanthogaster 3 0.001 64 63 70 64 63 69 70 69 126 | BAe 17 30 68 122 18 AMERICAN MUSEUM NOVITATES have been evident, or even looked for, in a widely distributed “‘biological species.”’ For example, distributional records show that S. gabonensis is narrowly endemic to the Cam- eroon-Gabon region’s humid coastal lowland forest. The restricted nature of its range, and possibly its ecology, were not fully appreci- ated when it was ranked as a subspecies. Recognizing these taxa as phylogenetic spe- cies may consequently have relevance for their conservation. If conservation priorities are predicated on knowledge of species ecol- ogies, and those ecologies are based on a bi- ological species concept, then there may be a risk that ecologies of target populations are being misunderstood. For example, if con- servation plans for a biological species called “Stiphrornis erythrothorax”’ were based on information from populations distributed somewhere other than these coastal forests, then the genetic, morphological and ecolog- ical distinction displayed by S. gabonensis could be misunderstood. The same may be true for S. sanghensis or any other polytypic biological species. The systematic perspective provided by Stiphrornis has broad implications. Our re- sults suggest that taxonomic variation may be underreported in maps in the Atlas of Hall and Moreau (1970) as a result of their appli- cation of the biological species concept. Thus, although the taxonomic schemes of Sclater (1930) and White (1962) allowed for geographic variation in Stiphrornis through the recognition of three subspecies, Hall and Moreau (1970) chose to map Stiphrornis as monotypic (although they discussed some variation in the accompanying text). Further systematic revisions of African lowland for- est birds may reveal other groups in which geographic and taxonomic variation have been obscured. For example, in a reanalysis of the phylogenetic relationships among the species of Bleda (Pycnonotidae) using mor- phometric and vocalization characters, Chap- puis and Erard (1993) found one subspecies to be more closely related to another biolog- ical species than to its purported conspecif- ics; these results, in conjunction with our own, indicate that phylogenetic and biogeo- graphic patterns within the African forest bi- ome cannot be recovered using the biological species concept. NO. 3270 These examples underscore the point that incomplete mapping of taxonomic variation has lead to misunderstandings about African biogeography, and as a consequence, inter- pretations (e.g., Diamond and Hamilton, 1980; Endler, 1982; Mayr and O’ Hara, 1986) based on the distributional data in Hall and Moreau (1970) may need to be reexamined. It could be argued, for example, that Endler’s (1982) failure to find contact zones evenly distributed between three purported refuges (areas of endemism) might be interpreted not as evidence against the refuge theory per se but as an indication that taxonomic patterns of variation were insufficiently resolved to reveal areas of contact accurately. Our poor knowledge about patterns of geographic variation also raises questions about the spatial units of biogeographic anal- ysis (areas of endemism). Moreau (1966) recognized a “‘Lower Guinea’”’ zoogeograph- ic region that included the Congo Basin west to southern Nigeria, but the distributional patterns as well as area relationships de- scribed here for the sister taxa Stiphrornis erythrothorax and S. gabonensis (Upper Guinea + Cameroon-Gabon) reinforce the composite nature of the southern Nigerian forest as noted by Marchant (1954). Similar- ly, Prigogine (1988) has suggested areas of endemism for birds in addition to those de- scribed by Diamond and Hamilton (1980). Additional phylogenetic analyses combined with better knowledge of taxon distributions should clarify the identity and composition of avifaunal areas of endemism. Although no biogeographic hypotheses have been proposed about historic relation- ships among African lowland forest areas of endemism, descriptions of processes under- lying African avifaunal diversity have usu- ally relied on vicariance through changes in forest cover caused by Pleistocene climatic fluctuations (e.g., Diamond and Hamilton, 1980; Louette, 1981; Prigogine, 1988). Few of these explanations have involved phylo- genetic analysis of relevant endemics. At the same time, little is presently known about the impact of climatic change on past forest dy- namics within the Guineo-Congolian forest (Maley, 1996). The location of Pleistocene lowland forest refuges has instead been in- ferred from areas of endemism for birds and 1999 mammals, although some Pleistocene forest contractions have recently been corroborated by paleoclimatic data in West Africa (e.g., Maley, 1996; Maley and Brenac, 1998). Whether or not the location of putative Pleistocene forest refugia will be corroborat- ed from paleobotanical data, the results of this analysis implicate Mt. Cameroon and its associated highlands as the zoogeographic barrier between S. gabonensis and S. ery- throthorax. Although a “‘molecular clock,”’ if in fact one exists, has not been calibrated for passerine birds, some workers accept a rate of 2.0% mitochondrial cytochrome D se- quence divergence per million years (see Klicka and Zink, 1997; Avise and Walker, 1998); under this rate assumption, the diver- gence of S. erythrothorax and S. gabonensis, as well as the split between (S. erythrothor- ax, S. gabonensis) and (S. sanghensis, S. xan- thogaster) may have occurred approximately 3 million years ago, during the Upper Plio- cene. These results and interpretations are at least suggestive that geomorphological and hydrological factors should be studied in ad- dition to paleoclimate for their possible role in vicariant events within the African low- land forests. Fjeldsa and Lovett (1997) proposed that most African lowland forest birds are phy- logenetically ‘‘old’”’ and predate the Pleisto- cene. In their model, based on relative branching rates derived from DNA-hybrid- ization phylogenies, lowland areas of high species richness are the result of the carrying capacity of specific habitats, phyletic speci- ation, and the immigration of “‘younger”’ species from topographically complex areas (in which diversity is created according to a disturbance-regime model) (Fjeldsa and Lov- ett, 1997). It is difficult to evaluate such complex causal models of speciation pro- cesses unless the patterns invoked are well understood, in terms of both geographic var- iation and phylogenetic relationships among the taxa involved (Cracraft, 1989; Bates et al., 1998). At this time it is apparent that patterns of variation of birds within the Guineo-Congo- lian forest are poorly known. The zoogeo- graphically complex nature of the north- western part of the Congo Basin forest that is emerging through our discovery of S. BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS 19 sanghensis as well as from surveys of other vertebrates (Fay, 1988; Joger, 1990) under- scores the need for more careful analyses of the distribution of geographic variation in and around the Congo Basin. Although new species are being described by other verte- brate zoologists (e.g., Joger, 1990) including a new species of shrew from the CAR (Ray and Hutterer, 1996), since 1966 no new avian species has been described from the Congo Basin forest (Mayr and Vuilleumier, 1983; Vuilleumier and Mayr, 1987; Vuilleumier et al., 1992; Hockey, 1997). This study, how- ever, suggests that this situation may be an artifact of viewing patterns of geographic variation and endemism through the lens of a particular species concept. Future work in African zoogeography, especially for low- land forest taxa, should focus on patterns of geographic variation within the framework of a phylogenetic species concept in order to discover differentiated taxa, their areas of en- demism, and the barriers responsible for them, as well as to test process-oriented hy- potheses relating to paleoclimatic and geo- logical events affecting biotic history. REFERENCES Avise, J. C., and D. Walker 1998. Pleistocene phylogeographic effects on avian populations and the speciation process. Proc. R. Soc. London. B 265: 457-463. Bates, J. M., S. J. Hackett, and J. Cracraft 1998. Area-relationships in the Neotropical lowlands: an hypothesis based on raw distributions of passerine birds. J. Bio- geogr. 25: 783-793. Brosset, A., and C. Erard 1986. Les oiseaux des régions forestiéres du nord-est du gabon. Vol. 1. Ecologie et comportement des espeéces. Paris: Soc. Natl. Prot. Nat. Bruhl, C. A. 1997. Flightless insects: a test case for histor- ical relationships of African mountains. J. Biogeogr. 24: 233-250. Chapin, J. P. 1953. The birds of the Belgian Congo, Pt. 3. Bull. Am. Mus. Nat. Hist. 75A. 1954. The birds of the Belgian Congo, Pt. 4. Ibid. 75B. Chappuis, C., and C. Erard 1993. Species limits in the genus Bleda Bon- 20 AMERICAN MUSEUM NOVITATES aparte, 1857. Z. zool. Syst. Evolutions- forsch. 31: 280-299. Charif, R. A., S. Mitchell, and C. W. Clark 1995. Canary 1.2 User’s manual. Ithaca, NY: Cornell Lab. Ornith. Cracraft, J. 1989. Speciation and its ontology: the empir- ical consequences of alternative species concepts for understanding patterns and processes of differentiation, pp. 28-59. In D. Otte and J. A. Endler (eds.), Spe- ciation and its consequences. Sunder- land, MA: Sinauer Assoc. Crowe, T. M., and A. A. Crowe 1982. Patterns of distribution, diversity and endemism in Afrotropical birds. J. Zool. (London) 198: 417-442. Desjardins, P., and R. Morais 1990. Sequence and gene orgainization of the chicken mitochondrial genome. J. Mol. Biol. 212: 599-634. Diamond, A. W., and A. C. Hamilton 1980. The distribution of forest passerine birds and Quaternary climatic change in tropical Africa. J. Zool. (London) 191: 379-402. Dowsett, R., and E Dowsett-Lemaire (eds.) 1997. Flore et faune du Parc National d’Odzala, Congo. Turaco Res. Rep. 6. Endler, J. A 1982. Pleistocene forest refuges: fact or fan- cy? In G. Prance (ed.) Biological di- versification in the tropics: 641-656. New York: Columbia Univ. Press. Fay, J. M. 1988. Forest monkey populations in the Cen- tral African Republic: the northern lim- its. A census in Manovo-Gounda-St. Floris National Park. Mammalia 52(1): 57-74. Fjeldsa, J., and J. Lovett 1997. Geographical patterns of old and young species in African forest biota: the sig- nificance of specific montane areas as evolutionary centres. Biodivers. Con- serv. 6: 325-346. Green, A. A., and R. W. Carroll 1991. The avifauna of Dzanga-Ndoki Nation- al Park and Dzanga-Sangha Rainforest Reserve, Central African Republic. Malimbus 13: 49-66. Hackett, S. J. 1996. Molecular phylogenetics and biogeog- raphy of tanagers in the genus Ram- Phocelus (Aves). Mol. Phylogenet. Evol. 5(2): 368-382. Hall, B. P, and R. E. Moreau 1970. An Atlas of Speciation in African Pas- NO. 3270 serine Birds. London: British Museum (Natural History). Hartlaub, G. 1855. Beschreibung einiger neuen, von Herrn H.S. Pel, hollandischem residenten an der Goldkuste, daselbst gesammelten vogelarten. J. Ornithol. 17: 344-360. Hockey, P. 1997. New birds in Africa. Africa Birds and Birding 2(1): 39-44. Irwin, M. P. S., and P. A. Clancey 1974. A re-appraisal of the generic relation- ships of some African forest-dwelling robins (Aves: Turdidae). Arnoldia (Rhodesia) 6(34): 1-19. Jackson, E 1910. Bull. Br. Ornithol. Club 25: 85. Jensen, E P. 1989. A review of some genera of African chats (Aves: Muscicapidae, Erithacini). Steenstrupia 15(6): 161-175. Joger, U. 1990. The herpetofauna of the Central Afri- can Republic, with a description of a new species of Rhinotyphlops (Serpen- tes: Typhlopidae). Jn G. Peters and R. Hutterer (eds.), Vertebrates in the trop- ics: 85-102. Bonn: Mus. Alexander Koenig. Keith, G. S., and W.W.H. Gunn 1971. Birds of the African rain forests. Sounds of nature 9. Fed. of Ontario Nat. and Am. Mus. Nat. Hist. Keith, S., E. K. Urban, and C. H. Fry 1992. The birds of Africa, vol. 4. London: Academic Press. Klicka, J., and R. Zink 1997. The importance of recent ice ages in speciation: a failed paradigm. Science 277: 1666-1669. Louette, M. 1981. The birds of Cameroon: an annotated checklist. Klasse Wet. (43) 163. 1984. Apparent range gaps in African forest birds. Proc. Pan Afr. Ornithol. Congr. 5: 275-286. 1992. Barriers, contact zones and subspecia- tion in central equatorial Africa. Bull. Br. Ornithol. Club 112A: 209-216. Maley, J. 1996. The African rain forest—main charac- teristics of changes in vegetation and climate from the Upper Cretaceous to the Quaternary. Proc. R. Soc. Edin- burgh B: 104: 31-74. Maley, J, and P. Brenac 1998. Vegetation dynamics, palaeoenviron- ments and climatic changes in the for- 1999 ests of western Cameroon during the last 28,000 years B.P. Rev. Palaeobot. Palynol. 99: 157-187. Marchant, S. 1954. The relationship of the southern Niger- ian avifauna to those of Upper and Lower Guinea. Ibis 96: 371-379. Mayr, E., and R. J. O’Hara 1986. The biogeographic evidence supporting the Pleistocene forest refuge hypothe- sis. Evol. 40(1): 55-67. Mayr, E., and R. A. Paynter, Jr. (eds.) 1964. Check-list of birds of the world, vol. 10. Cambridge, MA.: Mus. Comp. Zool. Mayr, E., and E Vuilleumier 1983. New species of birds described from 1966 to 1975. J. Ornithol. 124: 217-— 232, Moreau, R. E. 1966. The bird faunas of Africa and its is- lands. London: Academic Press. Prigogine, A. 1988. Speciation patterns of birds in the cen- tral African forest refugia and their re- lationship with other refugia. Acta XIX Congr. Int. Ornithol. 2: 144-157. Ot- tawa: Univ. Ottawa Press. Ray, J. C., and R. Hutterer 1996. Structure of a shrew community in the Central African Republic based on the analysis of carnivore scats, with the de- scription of a new Sylvisorex (Mam- malia: Soricidae). Ecotropica 1: 85—97. Reichenow, A. 1874. J. Ornithol.:103. Roy, M. S. 1997. Recent diversification in African green- buls (Pycnonotidae: Andropadus) sup- ports a montane speciation model. BERESFORD AND CRACRAFT: AFRICAN FOREST ROBINS pall Proc. R. Soc. (London). B 264: 1337— 1344. Sclater, W. L. 1930. Systema Avium Aethiopicarum Pt. 2. London: Br. Onithol. Union. Sibley, C. G., and B. L. Monroe, Jr. 1990. Distribution and Taxonomy of Birds of the World. New Haven, CT: Yale Uni- versity Press. Sharpe, R. B. 1883. Catalogue of the Passeriformes in the British Museum: Cichlomorphae Pt. 4. London: British Museum 1903. A hand-list of the genera and species of birds vol 4. London: British Museum (Natural History). Snow, D. 1978. An Atlas of Speciation in African Non- passerine Birds. London: British Mu- seum (Natural History). Swofford, D. L. 1998. PAUP*. Phylogenetic analysis using parsimony (*and other methods). Ver- sion 4. Sunderland, MA: Sinauer As- sociates. Vuilleumier, EF, and E. Mayr 1987. New species of birds described from 1976 to 1980. J. Ornithol., 28: 137— 150. Vuilleumier, EF, M. LeCroy , and E. Mayr 1992. New species of birds described from 1981 to 1990. Bull. Br. Ornithol. Club 112A: 267-309. White, C.M.N. 1962. A revised check-list of African shrikes, orioles, drongos, starlings, crows, wax- wings, cuckoo-shrikes, bulbuls, accen- tors, thrushes and babblers. Lusaka: Govt. Printer. White, EF 1983. The vegetation of Africa. Paris: UNESCO. 22 AMERICAN MUSEUM NOVITATES NO. 3270 APPENDIX 1 Gazetteer of Stiphrornis Specimens Examined erythrothorax LIBERIA Bong Co. 7°00'N 9°40'W Babeyru 33 N22 732'E Cape Mount 6°45'N 11°23'’W Basankusu 1°14'N 19°48'E Dugbe R. 4°51'N 8°46’W Bassin Lindi 0°33'S 28°40’E Gbarnga 7°02'N 9°26'W to 0°32'N 25°07E Grassfield 7°30'N 8°35'W Beritee 3°25'N 24°32'E Mt. Nimba 7°35'N 8°28'W Epulu 1°25'N 28°35'E Zigida 8°Q2'N 9°29'W Gamangui 210 N27 1S*E Ibanga 2°47'S/28°25'E ae Tbembo 2°38/N 23°37’E Sugar Loaf 8°25'N 13°14’W Ikela 1°11'S 23°16’E NrGeRIA Luhoho R. 4°30'S 35°08'E Degema 4°48'N 6°45'E . Caan age Ede 7°A0'N 4°30'E Kailo 2°38'S 26°07'E Fadom Kagomi 9°30'N 8°00'E Lukolela Poe's W710 E leat 6°55'N 3°00’E Lusambo 4°58'S 23°27 E Lagos 6°28'N 3°25’E Mauda 4°05'N 27°41’E Mamu Forest 6°10’N 7°10'E ee ioe Owerri 5°30'N 7°01'E ogee 2°21'S 26°49'E Ngayu 1°45'N 27°33/E gabonensis Panga 3°17'N 26°44'E CAMEROON Poko 3°09'N 26°53/E Dume 4°18'N 13°28'E ede Efulen 2°42'N 10°30’E Eseka 3°30'N 10°46'E Kipkabus 0°18'N 35°31'E Kribi 2°57'N 9°55'E UGANDA Kumba 4°38'N 9°25'E Budongo Forest 1°40'N 31°25'E Lolodorf 3°10'N 10°42’E to 1°53'N 31°41’E Mamfe S*46'N 9°17'E Katera 0°45'N 31°59’E Mbonge 4°35'N 9°05'E Luwala 0°27'N 33°07'E Melan 3°S UN 11°30'E Mabira Forest 0°23'N 32°54'E Nditam 5°21'N 11°14’E to 0°32'N 33°07’E Ndoi 4°5S'N 9°30'E Mutwanga 0°20'N 29°45'E Ngoumé S°30'N 11°26'E Ntandi 0°48'N 31°59’'E Victoria 4°00'N 9°12'E Semliki Valley 0°08'S 29°36'E GABON to 1°12’N 30°30’E Bassin D’Ivindo 0-2°N 12-14°E Eugalambo 0:24N 3300355: Bélinga 0°36'N 13°08'E CAR Kango O°10'N 10°09"E Ouessi River 5°35'N 24°36'E Kribi 2°56'N 9°56'E Oveng 2°25'N 12°16'E SUDAN Bangangai Forest 4°51'N 27°45’E EQ. GUINEA Bioko 3°13’N 8°24'E CAMEROON to 3°48'N 8°58’E Assobam 3°17'N 14°04’E xanthogaster Bitye 3°10'N 12°20'E DRC Mieri A°14'N 13°58’E Angumu 0°7'S 27°41’E Yokoudouma 3°25'N 15°03'E Avakubi 1°21'N 27°40'E